Data-Driven Decision-Making in Homelessness Services

Choice Modelling and Discrete Choice Experiments Data

Choice modelling, a flexible tool originating from market research, is commonly used in health services but less frequently in social care. It enables the measurement of preferences between alternative scenarios or types of service provision. In my research, I use discrete choice experiment (DCE) survey methods to collect information. 

Analysing DCE data unveils interesting insights: 

• What aspects of a service matter most to people, whether location of care and who provide care are important whereas how many visits of the carer per week is not prioritised. 
• Trade-offs between characteristics, such as how long they’re willing to wait for receiving care at home (compared with in temporary accommodation) or, with consideration of cost attributes, how much people are willing to pay extra for receiving care at home (compared with in temporary accommodation). 
• Willingness to pay for a policy change where multiple characteristics are changed, like how much people are willing to pay extra for Service A (where the main carer is a nurse, care is provided at home with biweekly visits) compared with Service B (where the main carer is a housing officer, care is provided in a temporary accommodation with one visit per week). 
• Uptake of different care models and use the DCE data to predict how may are likely to choose service A (85%) over Service B (15%). 

In simple terms, choice modelling and the analysis of DCEs data help understand what people want. This way, those in charge of service commissioning and delivery can make services better, meet specific needs, and ensure everyone gets the right help. Read more about the approach in this paper: ‘Applying discrete social experiments in social care research’ [PDF, 413KB] 

Applying DCEs to homelessness service delivery

With my NIHR SSCR fellowship, I aim to expand the application of choice modelling and the analysis of DCEs data, focussing on supporting their impact in social care homelessness. As part of a larger evaluation of the Department of Health and Social Care-funded National Out of Hospital Care Models programme for people who are homeless I am using the preference data collected to capture service users’ preferences for different types of specialist hospital discharge schemes for people who are homeless. Read more about my work applying DCE’s to homelessness here.  

As discussed in my webinar for KCL Homeless Webinar series (2023), DCE data visualisations present stated preferences data alongside other important information about service delivery costs and economic outcomes, patient follow and staff workload, health and housing outcomes, people sociodemographic and experience of care.  

These integrated management dashboards, created by my team, can support national decision-making for the health and wellbeing of the population. In alignment with NHS England, the future automatisation of the dashboards aims to facilitate rapid improvements in data quality and coverage, embedding real-time data into day-to-day operational practices at the local system, regional, and national levels. They also help service providers make a strong case for future funding to local commissioners.   

The lack of robust information on how homeless individuals value services highlights the importance of this work for service planners, commissioners and providers to plan, organize, and improve services effectively. Through the analysis of DCEs data, we gain valuable insights into service user preferences, trade-offs between characteristics, and willingness to pay for specific changes in care models.